Coronary heart disease caused 476,124 deaths in 1996 and continues to be the leading cause of death in America today. Over 12 million people alive today have a history of myocardial infarction, angina pectoris, or both. Every year in the US about 5 million perfusion studies are performed to evaluate extent and severity of CAD, thereby enabling clinical decisions regarding diagnosis, prognosis, and therapy for patients with heart disease. Of these, 1 million undergo angioplasty and about 500,000 have bypass surgery, and millions of others undergo drug therapy and or lifestyle changes to prevent progression of cardiac disease. It is widely recognized that computer quantification of myocardial perfusion images improves diagnostic accuracy and enhances confidence and reproducibility of interpretation. Theses quantitative approaches are well- established for assessing abnormalities in myocardial perfusion and function. However, they have not been developed or optimized for detecting changes in serial studies of the same patient such as is needed for assessing the effect of interventions, medical therapy, or disease progression. In this project, we will develop and validate computer-based methods to automatically quantify and visualize serial changes in myocardial perfusion and function from perfusion SPECT. The work can be separated into 4 projects: 1) To assess changes in myocardial perfusion, 2) to assess changes in myocardial function, 3) to design a virtual heart suitable for creating simulation data for optimizing and analyzing our algorithms, and 4) to validate the methods using both simulations and animal studies. Important subprojects include: a) development of 3-d and 4-d surface detection methods for defining LV endocardial and epicardial surfaces, b) development of algorithms for non-rigid alignment of static and/or dynamic serial SPECT images so that they may be more directly compared, c) development of motion analysis methods using similar non-linear alignment techniques to measure regional myocardial function and also to correct ungated SPECT scans for motion blur, and d) creation of new statistical approaches to determine significant changes in both global and regional perfusion and functional variables. The ultimate goal of this project is to create a clinically useful tool for detecting changes in serial SPECT studies. Most importantly, the tools will be extremely well characterized as to their sensitivity in detecting small changes as well as for overall accuracy.